The existing paradigm for the use of RNA interference (RNAi) in the development of small-RNA therapeutics and biologic tools is to interfere with the expression of a single gene using a short-hairpin-loop RNA (shRNA) or short-interfering RNA (siRNA). This existing single-gene-interference paradigm, which we seek to challenge, derives from the use of shRNAs or siRNAs as research tools to gain insights into the possible functions of proteins of interest. However, the most potent endogenous microRNAs (miRNAs), on which shRNAs and siRNAs are modeled, target hundreds of mRNAs simultaneously through seed-sequence matches of 7-8 nucleotides. Not surprisingly, therefore, efforts to develop small-RNA therapeutics and biologic tools based on interfering with the expression of single genes are plagued by off-target effects, which are likely to result in poor therapeutic indices. In addition, miRNAs have recently been shown to directly activate, as well as interfere with, gene expression. The rules governing this RNA activation (RNAa) are unknown. Indeed, the complete range of small-RNA targets and effects is only beginning to be appreciated. To harness the full potential for the development of small-RNA therapeutics and biologic tools, including multi-gene targeting and RNA activation, a fundamentally different approach is needed. With this in mind, we designed and synthesized the first shRNA-expressing library that is completely random at the nucleotide level. Cell-based screening assays using our library are unbiased with respect to mechanism(s) of action - in effect, we let the cells tell us which small RNAs are the most effective and least toxic. Hundreds of thousands of random shRNAs can be screened in a single tissue-culture dish using selection assays and a pooled approach. Because there are only approximately 20,000 possible seed sequences (for canonical RNAi), and because shRNAs are bio-active molecules, hit sequences are invariably present. Optimization of initial hit sequences, by random mutagenesis and re-screening, is straightforward. Our approach allows us to identify the most effective, and least toxic, small RNAs to be used as therapeutics or biologic tools. We propose to use our library to identify and optimize shRNA sequences for stem-cell induction and for cell differentiation. Library sequences that we identify and optimize could be expressed from vectors as shRNAs, or transfected into cells as siRNAs, which exert their effects without genomic integration. Reporter constructs for Nanog and Oct4 will be used to identify shRNAs for stem-cell induction, and reporter constructs for NKX2-5 and Ngn3 will be used to identify shRNAs for cardiac and beta-cell differentiation, respectively. Profiling by microarray and/or proteomic analysis will be used to identify unique target gene sets.